Rave: A Computational Framework to Facilitate Research in Design Decision Support

Abstract

The cognitive challenges in the design of complex engineered systems include the scale and scope of decision problems, nonlinearity of the trade space, subjectivity of the problem formulation, and the need for rapid decision making. These challenges have motivated an active area of research in design decision-support methods and the development of commercial and openly available design frameworks. Although these frameworks are extremely capable, most are limiting as a basis for research relating to design decision support because they offer little user flexibility for incorporating and evaluating new features or techniques. This paper describes Rave (www.rave.gatech.edu), a computational framework designed specifically as a research platform for design decision-support methods. Rave has been structured to be flexible and adaptable, handle data with systematic data structures and descriptive metadata, facilitate a wide spectrum of visualization types, provide features to enable user interactivity and linking of graphics, and incorporate surrogate modeling and optimization as enabling capabilities. This framework is envisioned to provide the research and industrial communities an easily expandable and customizable baseline capability to facilitate investigation of further design decision-support advancements.

Tradespace Exploration of Reconfigurability using Surrogate Points to Simplify Dominance Determination

“…The ATSV software developed at Penn State University provides a large suite of tools for visualizing multi-dimensional spaces with large populations of designs 21 . RAVE is a computational software for decision support that seeks to provide flexibility to the user in how the data is presented and manipulated 22 .…”

Information visualization for selection in Design by Shopping

“…Each leaf of the tree specifies the expected output value (with color scale), as a consequence of the particular input values described by the path from the root to that leaf (Fig 4). The Rave tool (Daskilewicz & German, 2012) is a computational framework designed specifically as a research platform for design decision-support methods. It provides a "blank canvas" on which we can arrange graphs, tables, images, text, and user interface controls (automatically linked together).…”

“…There is already much research in visual design and it has been shown that fast graphical design interfaces impact user performance in terms of design efficiency, design effectiveness and the design search process (Ligetti et al, 2003). Various tools for exploring the design space with different graphs exist, namely: the ARL Trade Space Visualizer (Stump et al, 2004), the VIDEO tool (Kollat & Reed, 2007), the LIVE tool (Yan et al, 2011(Yan et al, , 2012) and the Rave tool (Daskilewicz & German, 2012).…”

Exploring the Effectiveness of Using Graveyard Data When Generating Design Alternatives

“…In multiobjective optimization problems, a designer must navigate the tradeoffs between competing objectives [1][2][3] and select the location on the Pareto frontier that best satisfies the preference structure of the designer. Similarly, tradespace analysis involves executing broad trade studies to understand which designs meet minimum requirements while locating those designs that best meet the goal of the project [4][5][6][7]. To aid in this selection process, design approaches have been created that capture the preference structure of the designer using hypothetical alternatives [8,9] or reduce the problem to a single metric using value or utility functions [10][11][12].…”

“…The first objective (f 1 ) of this case study is to minimize the cross-sectional area as seen in Eq. (7), while the second objective (f 2 ) is to minimize the vertical deflection of the beam subject to loads P and Q as seen in Eq. (8).…”